Draft version April 14, 2020

Typeset using LATEX preprint style in AASTeX63

Simultaneous X-ray and Radio Observations of the Repeating Fast Radio BurstFRB 180916.J0158+65

P. Scholz,1 A. Cook,1, 2 M. Cruces,3 J. W. T. Hessels,4, 5 V. M. Kaspi,6, 7 W. A. Majid,8, 9

A. Naidu,6, 7 A. B. Pearlman,9, ∗ L. Spitler,3 K. M. Bandura,10, 11 M. Bhardwaj,6, 7

T. Cassanelli,1, 2 P. Chawla,6, 7 B. M. Gaensler,1, 2 D. C. Good,12 A. Josephy,6, 7

R. Karuppusamy,3 A. Keimpema,13 A. Yu. Kirichenko,14, 15 F. Kirsten,16 J. Kocz,9

C. Leung,17, 18 B. Marcote,19 K. Masui,17, 18 J. Mena-Parra,17 M. Merryfield,6, 7

D. Michilli,6, 7 C. J. Naudet,8 K. Nimmo,4, 5 Z. Pleunis,6, 7 T. A. Prince,9, 8

M. Rafiei-Ravandi,20 M. Rahman,1 K. Shin,17, 18 K. M. Smith,20 I. H. Stairs,12

S. P. Tendulkar,6, 7 and K. Vanderlinde1, 2

1Dunlap Institute for Astronomy & Astrophysics, University of Toronto, 50 St. George Street, Toronto, ON M5S 3H4,Canada

2David A. Dunlap Institute Department of Astronomy & Astrophysics, University of Toronto, 50 St. George Street,Toronto, ON M5S 3H4, Canada

3Max Planck Institut fur Radioastronomie, Auf dem Hugel 69, D-53121, Bonn, Germany4Anton Pannekoek Institute for Astronomy, University of Amsterdam, Science Park 904, 1098 XH, Amsterdam, The

Netherlands5ASTRON, Netherlands Institute for Radio Astronomy, Oude Hoogeveensedijk 4, 7991 PD Dwingeloo, The

Netherlands6Department of Physics, McGill University, 3600 rue University, Montreal, QC H3A 2T8, Canada7McGill Space Institute, McGill University, 3550 rue University, Montreal, QC H3A 2A7, Canada

8Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA 91109, USA9Division of Physics, Mathematics, and Astronomy, California Institute of Technology, Pasadena, CA 91125, USA

10CSEE, West Virginia University, Morgantown, WV 26505, USA11Center for Gravitational Waves and Cosmology, West Virginia University, Morgantown, WV 26505, USA

12Department of Physics & Astronomy, 6224 Agricultural Road, Vancouver, BC V6T 1Z1, Canada13Joint Institute for VLBI ERIC, Oude Hoogeveensedijk 4, 7991PD Dwingeloo, The Netherlands

14Instituto de Astronoma, Universidad Nacional Autnoma de Mxico, Apdo. Postal 877, Ensenada, Baja California22800, Mxico

15Ioffe Institute, 26 Politekhnicheskaya st., St. Petersburg 194021, Russia16Department of Space, Earth and Environment, Chalmers University of Technology, Onsala Space Observatory, 439

92, Onsala, Sweden17MIT Kavli Institute for Astrophysics and Space Research, Massachusetts Institute of Technology, 77 Massachusetts

Ave, Cambridge, MA 02139, USA18Department of Physics, Massachusetts Institute of Technology, 77 Massachusetts Ave, Cambridge, MA 02139, USA

19Joint Institute for VLBI ERIC, Oude Hoogeveensedijk 4, 7991 PD Dwingeloo, The Netherlands20Perimeter Institute for Theoretical Physics, 31 Caroline Street N, Waterloo ON N2L 2Y5, Canada

ABSTRACT

Corresponding author: P. Scholzpaul.scholz@dunlap.utoronto.ca

arX

iv:2

004.

0608

2v1

[as

tro-

ph.H

E]

13

Apr

202

0

2

We report on simultaneous radio and X-ray observations of the repeating fast radioburst source FRB 180916.J0158+65 using the Canadian Hydrogen Intensity MappingExperiment (CHIME), Effelsberg, and Deep Space Network (DSS-14 and DSS-63) ra-dio telescopes and the Chandra X-ray Observatory. During 33 ks of Chandra obser-vations, we detect no radio bursts in overlapping Effelsberg or Deep Space Networkobservations and a single radio burst during CHIME/FRB source transits. We detectno X-ray events in excess of the background during the Chandra observations. Thesenon-detections imply a 5-σ limit of < 5 × 10−10 erg cm−2 for the 0.5–10 keV fluence ofprompt emission at the time of the radio burst and 1.3 × 10−9 erg cm−2 at any timeduring the Chandra observations at the position of FRB 180916.J0158+65. Given thehost-galaxy redshift of FRB 180916.J0158+65 (z ∼ 0.034), these correspond to energylimits of < 1.6 × 1045 erg and < 4 × 1045 erg, respectively. We also place a 5-σ limitof < 8 × 10−15 erg s−1 cm−2 on the 0.5–10 keV absorbed flux of a persistent sourceat the location of FRB 180916.J0158+65. This corresponds to a luminosity limit of< 2 × 1040 erg s−1. Using Fermi/GBM data we search for prompt gamma-ray emis-sion at the time of radio bursts from FRB 180916.J0158+65 and find no significantbursts, placing a limit of 4 × 10−9 erg cm−2 on the 10–100 keV fluence. We also searchFermi/LAT data for periodic modulation of the gamma-ray brightness at the 16.35-dayperiod of radio-burst activity and detect no significant modulation. We compare thesedeep limits to the predictions of various fast radio burst models, but conclude that sim-ilar X-ray constraints on a closer fast radio burst source would be needed to stronglyconstrain theory.

Keywords: X-rays: bursts, X-rays: general, gamma rays: general, stars: neutron

1. INTRODUCTION

Fast radio bursts (FRBs) are a new class of radio transient with unknown origins (see Cordes& Chatterjee 2019; Petroff et al. 2019, for reviews). They are millisecond-long, bright (peak fluxdensities ∼ 0.1–10 Jy at ∼ 1 GHz) bursts and have been observed at frequencies from 300 MHz(Chawla et al. 2020) to 8 GHz (Gajjar et al. 2018). Their distances, both based on their dispersionmeasure (DM) excesses (in comparison to the expected Milky Way contributions; Cordes & Lazio2002; Yao et al. 2017) and measured host-galaxy redshifts for a few sources (Chatterjee et al. 2017;Bannister et al. 2019; Ravi et al. 2019; Prochaska et al. 2019; Marcote et al. 2020), are extragalactic,and the most distant sources appear to come from cosmological distances (i.e., z >∼ 0.5; Thornton et al.2013). The extreme luminosities and short duration of FRBs point to coherent emission originatingfrom a compact object. Prior to the discovery of repeat bursts from some FRB sources, most modelsinvoked cataclysmic phenomena to explain the extreme energetics of FRBs (for a catalog of models,see Platts et al. 2018). However, since the discovery of repeat bursts from FRB 121102 (Spitler et al.2016), models that can account for repetition have become increasingly the focus of theoretical work.

One central engine in particular has garnered a lot of attention: the millisecond magnetar. In thismodel, an FRB is powered by a young, recently formed millisecond magnetar (e.g., Lyubarsky 2014;

∗ NDSEG Research Fellow.NSF Graduate Research Fellow.

3

Beloborodov 2017; Metzger et al. 2017) and may have a high-energy counterpart. The older, muchless energetic, magnetars in our Galaxy are known to power X-ray and gamma-ray bursts and flareson timescales of milliseconds to seconds (see Kaspi & Beloborodov 2017, for a review), which aresimilar to the duration of FRBs. The high-energy burst emission of magnetars comes in at least twoclasses: giant flares and short X-ray bursts. To date, only three magnetar giant flares have beendetected in our Galaxy (Evans et al. 1980; Hurley et al. 1999, 2005) with X-ray peak luminosities inthe range ∼ 1044−1047 erg s−1. Short X-ray bursts from magnetars are emitted much more frequentlybut are much fainter than giant flares (peak X-ray luminosities of ∼ 1036 − 1043 erg s−1; e.g., Goguset al. 1999, 2000; Scholz & Kaspi 2011).

Scholz et al. (2017) undertook several campaigns of coordinated X-ray and radio observations ofFRB 121102, to probe for coincident high-energy emission during the radio bursts. With theseobservations, upper limits were placed on X-ray (0.5–10 keV) and gamma-ray (10–100 keV) emissionat the time of radio bursts. Owing to the relatively large distance to FRB 121102 (z ∼ 0.193;luminosity distance of 972 Mpc), these limits were found to be ∼ 10× above what is expected for amagnetar giant flare (Scholz et al. 2017).

The recent success of the CHIME/FRB Collaboration in discovering repeating FRBs has ledto several sources that could be much closer than FRB 121102, based on their low DM excesses(CHIME/FRB Collaboration et al. 2019). One of these sources, FRB 180916.J0158+65, was subse-quently localized with milliarcsecond precision to a spiral galaxy at z = 0.0337 ± 0.0002 (luminositydistance of 149 Mpc) using observations from the European VLBI Network (Marcote et al. 2020).Recently, a 16.35-day periodicity in the burst activity of FRB 180916.J0158+65 was found, where thesource seems to be active in a ∼ 5 day window (CHIME/FRB Collaboration et al. 2020), although analiased, shorter period cannot presently be excluded. Armed with this localization, and knowledge ofthe periodic activity level, we were able to perform a deep, targeted, search for X-ray emission usingthe Chandra X-ray Observatory coordinated with radio observations at times when the detection ofradio bursts from the source were highly probable. The greater proximity of FRB 180916.J0158+65compared to FRB 121102 allows us to probe ∼ 40× deeper in energy for such emission. Previously,limits have been placed on the high-energy emission of FRB 180916.J0158+65 during its active phasesusing INTEGRAL (Panessa et al. 2020), Swift/XRT (Tavani et al. 2020a), and Chandra (Kong et al.2020)1 Other studies have also placed limits on the gamma-ray emission of a large sample of FRBsources (e.g., Tendulkar et al. 2016; Cunningham et al. 2019).

Here we present simultaneous deep X-ray and radio observations on 2019 December 3 and 18performed with the goal of detecting or constraining any X-ray counterparts to the radio burstsfrom FRB 180916.J0158+65. We also present a search for gamma-ray emission at the times ofradio bursts from FRB 180916.J0158+65. We describe the Chandra (X-ray), Fermi (gamma-ray),Effelsberg, Deep Space Network, and CHIME (radio) observations in Section 2. In Section 3 wepresent the results of our search for bursts in the radio observations as well as X-ray (Chandra) andgamma-ray (Fermi) emission both at the time of radio bursts and at anytime during the high-energyobservations. We discuss the significance of these results in Section 4.

2. OBSERVATIONS

1 based on the same Chandra observations presented here.

4

04:00 08:00 12:00 16:00 20:00 00:00 04:00 08:00

CHIME/FRBEffelsberg

DSS-14DSS-63

Chandra

2019 December 02

04:00 08:00 12:00 16:00 20:00 00:00 04:00UTC Time

CHIME/FRBEffelsberg

DSS-14DSS-63

Chandra

2019 December 18

Figure 1. Timeline of Chandra observations (purple) and the coordinated radio observations from CHIME/FRB

(red), Effelsberg (orange), and the Deep Space Network (DSS-14, in green, and DSS-63, in blue, telescopes). The bars

show the times when each telescope was observing FRB 180916.J0158+65. The arrow on 2019 December 18 marks

the time of the CHIME/FRB-detected burst.

2.1. Chandra X-ray Observatory

FRB 180916.J0158+65 was observed by Chandra on 2019 December 3 (ObsID 23081) and 2019December 18 (ObsID 23082) at epochs consistent with the “on-phase” of the periodic activity ofFRB 180916.J0158+65 identified by CHIME/FRB Collaboration et al. (2020). The ACIS-S3 detectorwas operated in VFAINT mode with a 1/8 sub-array read out providing a 8′ × 1′ field of view anda 0.4-s frame time. The exposure times were both ∼ 16 ks, as listed in Table 1. Figure 1 shows atimeline of the Chandra observations and how they overlap with radio observations.

The resulting data were analyzed using CIAO2 version 4.12 (Fruscione et al. 2006) following stan-dard procedures recommended by the Chandra X-ray Center. Source events were extracted froma 1′′-radius region (95% encircled energy) centered on the position of FRB 180916.J0158+65 andarrival times were corrected to the Solar-System Barycenter using the source position measured byMarcote et al. (2020) to a precision of ∼ 2 milliarcseconds with the European VLBI Network (EVN).

2.2. CHIME/FRB

The CHIME/FRB backend continuously searches total-intensity, polarization-summed time seriesfrom each of the 1,024 beams formed across CHIME’s 2◦× 120◦ field-of-view. The time series have a0.98304-ms time resolution and 16,384 frequency channels across the 400–800 MHz band. The back-end uses real-time radio-frequency interference (RFI) mitigation and a tree dedispersion algorithm tosearch over a wide range of trial DMs. Dispersed signals with integrated S/N values greater than thesystem’s configurable threshold are forwarded to a post-detection pipeline to classify sources as RFI,known Galactic sources, or unknown Galactic or extragalactic signals (by comparing to predictedGalactic contributions to DM). Signals are classified as FRBs (i.e., unknown extragalactic) if theyare not associated with any known Galactic sources, and their observed DMs exceed the maximum

2 Chandra Interactive Analysis of Observations. http://cxc.harvard.edu/ciao/

5

Table 1. Summary of Joint X-ray/Radio Observations

Telescope Obs ID/ Start time End time Exposure time

Frequency (MHz) (UTC) (UTC) (s)

Chandra23081 2019-12-03 01:33:03 2019-12-03 07:01:53 16390

23082 2019-12-18 03:47:29 2019-12-18 09:13:36 16300

CHIME/FRBa 400–800 2019-12-18 04:06:36 2019-12-18 04:21:22 886

Effelsberg 1210–15102019-12-02 23:29:27 2019-12-03 07:29:27 28800

2019-12-18 02:21:17 2019-12-18 09:21:17 25200

DSS-14 1360–1720

2019-12-02 05:37:28 2019-12-02 07:01:28 5040

2019-12-02 07:09:02 2019-12-02 07:50:16 2474

2019-12-18 02:22:04 2019-12-18 03:51:04 5340

2019-12-18 04:01:12 2019-12-18 05:35:12 5640

2019-12-18 06:05:42 2019-12-18 07:22:42 4620

2019-12-18 07:30:16 2019-12-18 09:04:16 5640

2019-12-18 09:11:46 2019-12-18 10:45:46 5640

2019-12-18 10:53:32 2019-12-18 12:18:32 5100

DSS-63

2019-12-02 02:10:34 2019-12-02 03:49:34 5940

2019-12-02 22:58:40 2019-12-03 00:37:40 5940

2205–2310 &2019-12-03 02:32:46 2019-12-03 03:43:29 4243

8180–85752019-12-18 20:53:48 2019-12-18 22:27:48 5640

2019-12-18 22:36:30 2019-12-19 00:10:30 5640

2019-12-19 00:18:26 2019-12-19 01:52:26 5640

2019-12-19 02:00:34 2019-12-19 03:43:14 6120aStart, end, and exposure times based on time spent by source within the 600 MHz FWHM of theCHIME/FRB formed beams.

values predicted by Galactic DM models (Cordes & Lazio 2002; Yao et al. 2017). See CHIME/FRBCollaboration et al. (2018) for a detailed description of the CHIME/FRB system.

On 2019 December 3, CHIME was offline for upgrades and so was unable to search for bursts atthat time. On 2019 December 18, FRB 180916.J0158+65 was within the FWHM (at 600 MHz) ofthe CHIME/FRB beams for 14.7 min (see Table 1). During that period, the source moved throughthe four columns of synthesized beams. As such, the sensitivity to FRB 180916.J0158+65 variedsignificantly over the course of the transit.

2.3. Effelsberg Radio Telescope

The Effelsberg 100-m radio telescope observed FRB 180916.J0158+65 with the 7-beam receiver(P217mm) at a center frequency of 1.36 GHz. The central beam was pointed at the precise positionmeasured by the EVN localization (Marcote et al. 2020). These observations spanned the full extent

6

of the Chandra observations (see Figure 1) and their start times, end times, and total on-source timeare given in Table 1. The PFFTS digital backend recorded total intensity spectral data with a timeresolution of 54.6 µs, 512 frequency channels, and a bandwidth of 300 MHz (∆ν = 0.586 MHz).Before processing, the PFFTS data were converted from 32-bit floats to 8-bit unsigned integers insigproc filterbank format.

The data were searched using the PRESTO search software (Ransom 2011)3. Broadband, impulsiveRFI was removed using an algorithm that first re-scales each frequency channel according to thestandard deviation and median of that channel and then calculates a zero-DM timeseries. Statis-tically anomalous time samples were identified by applying an S/N threshold, and values for eachfrequency channel in that time sample were replaced with Gaussian noise with the statistics of thatchannel. The cleaned filterbank was then passed to rfifind for further RFI excision. The datawere then downsampled by a factor of eight in time and dedispersed with 100 trial DMs rangingfrom 300 pc cm−3 to 400 pc cm−3 (FRB 180916.J0158+65 has a DM of 349 pc cm−3; CHIME/FRBCollaboration et al. 2019) in steps of 1 pc cm−3 with prepsubband. Each time series was convolvedwith a template bank of boxcar matched filters yielding effective time resolutions of 0.44 ms to 17.5ms, and candidate bursts were identified in each timeseries by applying a detection threshold ofS/N > 6 (single pulse search.py). The results were inspected by eye, and promising candidateswere further investigated by looking at a time-frequency snapshot around each candidate.

2.4. Deep Space Network

The Deep Space Network (DSN) observed FRB 180916.J0158+65 for a total of ∼ 22 hr, partiallyoverlapping with the Chandra observations (see Figure 1), using DSS-14 and DSS-63, two 70-mdiameter radio antennas located in Goldstone, California and Robledo, Spain. FRB 180916.J0158+65was observed at L-band (center frequency of 1.5 GHz; data recorded in left circular polarization)using DSS-14 for a total of 11 hr over eight separate scans. DSS-63 observed FRB 180916.J0158+65simultaneously at S-band (center frequency of 2.3 GHz) and X-band (center frequency of 8.4 GHz)with data recorded in both left and right circular polarization for a total of 11 hr in seven separatescans (see Table 1). The L-band system on DSS-14 spans roughly 500 MHz of bandwidth, but only250 MHz of the total bandwidth was usable during our observations after RFI mitigation. The dataat S-band and X-band were recorded with bandwidths of 105 MHz and 395 MHz, respectively.

Data were recorded using pulsar backends that record channelized power spectral density measure-ments in filterbank format. The L-band data were recorded with a time and frequency resolutionof 102.4µs and 0.625 MHz, respectively. The S-band and X-band data were recorded with a timeand frequency resolution of 2.2 ms and 0.464 MHz, respectively. We performed short observations ofa bright pulsar (PSR B0329+54) at various times throughout the observing campaign to validatethe quality of the data. The data were flux calibrated by measuring the Tsys at each frequency bandwhile the antenna was in the stow position. We then corrected the Tsys values for elevation effects,which were minimal since all of our observations occurred when the source elevation was above 20◦.

The data processing procedures followed those described in previous DSN studies of pulsars(e.g., Majid et al. 2017; Pearlman et al. 2018, 2019). In each data set, we corrected for the bandpassslope across the frequency band. Bad frequency channels corrupted by RFI were identified using thePSRCHIVE software (Hotan et al. 2004) and masked. We also subtracted the moving average from

3 https://github.com/scottransom/presto

7

each data point using 0.5 s around each time sample in order to remove any long timescale temporalvariability. The cleaned data from each epoch were then dedispersed with trial DMs between 300and 400 pc cm−3. We searched for FRBs using a matched filtering algorithm, where each dedispersedtime-series was convolved with logarithmically spaced boxcar functions with widths ranging between1–300 times the native time resolution. FRB candidates with detection S/N>6 were saved and clas-sified using a GPU-accelerated machine learning pipeline based on the FETCH (Fast ExtragalacticTransient Candidate Hunter) package (Agarwal et al. 2019).

2.5. Fermi Gamma Ray Space Telescope

The Fermi telescope has two sets of detectors on board, the Gamma-ray Burst Monitor (GBM;Meegan et al. 2009) and the Large Area Telescope (LAT; Atwood et al. 2009). The GBM consistsof 12 sodium iodide (NaI; 8 keV – 40 MeV) and 2 bismuth germanate (BGO; 300 keV – 40 MeV)scintillators pointed in various directions to provide all-sky coverage to gamma-rays. In this work weuse only the NaI detectors. The GBM instrument records data in several different data products,but here we use only the time-tagged events (TTE) data which provides event data with 2-µs timeresolution and 128 energy channels. The LAT is a pair-conversion telescope providing sensitivity togamma-ray photons in the range 20 MeV–300 GeV in a 2.4 sr (20% of sky) field of view. The LATimages the sky with a time resolution of 10 µs or better. The LAT collaboration periodically releasesimproved reprocessing of their gamma-ray events. Here we use the most recent release, Pass 8.

3. ANALYSIS AND RESULTS

3.1. Radio Bursts

During the 2019 December 18 transit of FRB 180916.J0158+65 over CHIME, which was simulta-neous with a Chandra observation, a single radio burst was detected by CHIME/FRB. The burstwas detected at MJD 58835.17721035 (barycentric after correcting for dispersive delay), 446 s afterthe start of the Chandra observation, with a band-averaged S/N of 12.8 which corresponds to a peakflux density of 0.4±0.2 Jy and fluence of 2.9±0.7 Jy ms (see CHIME/FRB Collaboration et al. 2020,for additional details on this burst).

In the simultaneous Effelsberg observations, no bursts with S/N > 6 were identified by the PRESTO

search. Assuming a system equivalent flux density of 20 Jy for the P217mm receiver and S/N > 6,the fluence threshold is 0.15 Jy ms

√(w/1 ms), where w is the burst duration in ms. The Effelsberg

time series were also manually inspected around the time of detected CHIME/FRB bursts and noexcess was found. In the DSN observations listed in Table 1, no radio bursts were detected. For apulse width of w, the fluence thresholds (for S/N > 6) on the peak flux densities during these epochsare: 0.25 Jy ms

√(w/1 ms) at L-band, 0.29 Jy ms

√(w/1 ms) at S-band, 0.14 Jy ms

√(w/1 ms)

at X-band.

3.2. Limits on Prompt X-ray Emission

We searched the Chandra observations both for X-ray photons arriving nearby in time to theCHIME/FRB-detected radio burst and at anytime during the observations. In the 2019 December 3Chandra observation, a single photon was detected at the source position, but there were no detectedradio bursts in overlapping radio observations. In the 2019 December 18 Chandra observation, asingle photon was detected at the source position, 4.7 hr after the CHIME/FRB-detected radio burstand 500 s before the end of the Chandra observation. We take into account the dispersion delay of the

8

radio bursts (9 s at 400 MHz) when comparing to the times of high-energy photons. The backgroundcount rate in the source extraction region during the observation was 6× 10−5 counts s−1. This leadsto a probability of 64% of detecting one or more photons within 4.7 hr of the radio burst. Given thishigh false alarm probability, we have no reason to associate the detection with FRB 180916.J0158+65.For both observations, the detection of a single X-ray count within the source extraction region ofFRB 180916.J0158+65 is consistent with the background count rate.

Following Scholz et al. (2017), we place upper limits using Poisson statistics and the Bayesianmethod of Kraft et al. (1991). For all limits in this work, we use a stringent confidence level of0.9999994, the equivalent of the 5-σ width of a Gaussian distribution. For brevity, we refer to thisconfidence level as “5-σ” below. We first derive a “model-independent” limit, that is, assumingan equal probability of a source photon occurring across the 0.5–10 keV band (note that this iseffectively assuming a flat spectral model with zero X-ray absorption; see below for exploration ofmore reasonable models). This 5-σ confidence upper limit on the 0.5–10 keV fluence for a single X-rayburst at the time of the detected radio bursts is 5×10−10 erg cm−2 corresponding to 1.6×1045 erg at theluminosity distance of FRB 180916.J0158+65. These fluence and energy limits are valid for any burstduration contained within 446 s before the radio bursts (i.e., from the beginning of the observation)and 4.7 hr after (i.e., up to the time of the Chandra background photon). The fluence limit for anX-ray burst arriving at any other time during the Chandra observations is 1.3× 10−9 erg cm−2 for anassumed duration of 5 ms, corresponding to an energy limit of 4 × 1045 erg.

As discussed in Scholz et al. (2017), the implied limit on the emitted energy of a putative X-rayburst depends strongly on the underlying spectral model of the burst. By assuming a spectral modeland taking into account the spectral response of Chandra, a fluence limit for that underlying spectralmodel can be calculated. To generate the assumed source spectra we used XSPEC v12.10.1f withabundances from Wilms et al. (2000) and photoelectric cross-sections from Verner et al. (1996). Inorder to enable direct comparison, we assume the same fiducial models used by Scholz et al. (2017)for FRB 121102: a blackbody spectrum with kT = 10 keV as observed in magnetar hard X-raybursts (e.g., Lin et al. 2012; An et al. 2014), a cutoff power-law with index Γ = 0.5 and cutoff energyof 500 keV, similar to a SGR 1806−20-like giant flare spectrum (Mazets et al. 2005; Palmer et al.2005) and a power-law model with index Γ = 2 as an example soft spectrum, a contrast to the hardmagnetar burst models. In Table 2 we show the resulting fluence and energy limits assuming thesesource models. For X-ray absorption, we assume two values, 1022 cm−2 and 1024 cm−2. The firstis a typical value for a sightline passing through the Milky Way and the disk of a Milky-Way-likehost galaxy and the second is an extreme value to show the effects of a high degree of absorptionfrom material close to the source (such as the surrounding supernova ejecta in the magnetar model;Metzger et al. 2017).

3.3. Limits on Persistent X-ray Emission

To place the best-possible limits on a persistent source we combined the two Chandra observationsfor a total of 33 ks of exposure time. In these two observations, only two events were detected ina 1′′-radius region centered on the position of FRB 180916.J0158+65. We measure a 0.5–10 keVbackground count rate in a 25′′-radius region chosen to be away from the source of 0.7 counts s−1

sq. arcsec−2. Given this background rate, the two detected counts are consistent with the backgroundin the combined observations. Using these detected and measured background rates, we measure a 5-σ count rate limit of 5.5×10−4 counts s−1, using the Bayesian method of Kraft et al. (1991). Assuming

9

Table 2. Burst limits from Chandra for different X-ray spectral models

Model NH kT/Γ Absorbed 0.5–10 keV Unabsorbed 0.5–10 keV Extrapolated 10 keV–1 MeV

( cm−2) (keV/-) Fluence Limit Energy Limita Energy Limita

(10−11 erg cm−2) (1045 erg) (1047 erg)

Blackbody 1022 10 90 3 0.7

Blackbody 1024 10 200 20 7b

Cutoff PL 1022 0.5 50 1.4 5

Cutoff PL 1024 0.5 180 20 90b

Soft PL 1022 2 20 0.9 0.014

Soft PL 1024 2 120 50 0.8aAssuming the measured luminosity distance to FRB 180916.J0158+65, 149 Mpc (Marcote et al. 2020).

bMore stringent limits on these models are available from Fermi/GBM. See Section 3.4.

Note—5-σ confidence upper limits. See Section 3.2 for details.

a photoelectrically absorbed power-law source spectrum with Γ = 2 and NH ∼ 1 × 1022 cm−2, the5-σ upper limit on the persistent 0.5–10 keV X-ray absorbed flux from FRB 180916.J0158+65 or itshost galaxy is 8 × 10−15 erg cm−2 s−1. At the luminosity distance of FRB 180916.J0158+65 thiscorresponds to an isotropic luminosity limit of 2 × 1040 erg s−1.

3.4. Limits on Prompt Gamma-ray Emission

We searched data from the Fermi/GBM for gamma-ray counterparts at the time of radio burstsfrom FRB 180916.J0158+65 using a similar analysis to that in Scholz et al. (2017). We searched theTTE GBM data in the energy range 10–100 keV for NaI detectors that were pointed < 60◦ from thesource position. The 2018 December 18 bursts in this work were not visible to GBM as the sourcewas occulted by the Earth at the time. However, of the 28 bursts in CHIME/FRB Collaborationet al. (2020), 12 bursts occurred at a time when TTE data were available and the source was < 60◦

from at least one NaI detector and not occulted by the Earth. For these bursts, we searched eachTTE timeseries for excess counts in 1- and 5-ms bins in a 20-s window centered on the arrival time ofthe CHIME/FRB detected radio burst (after correcting for the dispersive delay). We find no signalsthat are not attributable to Poisson fluctuations from the background count rate at a 5-σ confidencelevel. Taking into account the effective area of the NaI detectors4 towards the source position at thetime of each event, the background count rate, and assuming a burst timescale of 0.1 s, we place anupper limit of 2 × 10−8 erg cm−2 on the 10–100 keV fluence. This corresponds to a 10–100 keV burstisotropic energy limit of 6 × 1046 erg at the measured luminosity distance of FRB 180916.J0158+65.If we assume a burst of gamma-rays is emitted at the time of each radio burst, the limit becomes4 × 10−9 erg cm−2. At the measured luminosity distance of FRB 180916.J0158+65, this correspondsto a 10–100 keV burst energy limit of 1 × 1046 erg. These limits are more constraining than the

4 Generated using the GBM Response Generator https://fermi.gsfc.nasa.gov/ssc/data/analysis/rmfit/DOCUMENTATION.html

10

extrapolated limits for prompt emission from the Chandra observations presented in Table 2 for thehighly-absorbed hard (10 keV blackbody and cut-off power-law) models. For those fiducial modelsthe 10 keV to 1 MeV energy limits are 7× 1046 erg and 1× 1048 erg, respectively. No bursts from thiswork or CHIME/FRB Collaboration et al. (2020) occurred in the Fermi/LAT field-of-view.

3.5. Search for Periodic Gamma-ray Emission

All Fermi/LAT photons with energies above 1 GeV and within a 5◦ radius region around thecoordinates of the source were selected, conservatively reflecting the ∼ 3◦ 95% containment radiusfor the point spread function at 1 GeV. We then filtered the data based on event class and zenithangle to ensure data quality and exclude Earth-limb photons. This data spans all 11 years fromMJD 54683 to MJD 58907. We removed data outside of the Good Time Intervals and corrected forexposure in each phase bin, before folding the data at the measured 16.35-day period (CHIME/FRBCollaboration et al. 2020). We performed an H-test (de Jager et al. 1989) on the resultant pulseprofile and find no significant signal, with a false-alarm probability of 31.3%.

4. DISCUSSION

4.1. Comparison to Previous Limits

The limits determined here can be compared to the similar campaign performed for FRB 121102using XMM-Newton and Chandra observations that were simultaneous with radio observations(Scholz et al. 2017). Figure 2 shows the limits, in burst energy, as a function of photon energyfor both FRB 121102, from Scholz et al. (2017), and FRB 180916.J0158+65, from this work. AsFRB 180916.J0158+65 is 6.5 times closer than FRB 121102, the single-burst energy limits fromChandra ACIS and Fermi GBM observations are ∼ 40× more constraining. However, the campaignon FRB 121102 included several Chandra and XMM-Newton observations during which 11 radiobursts were detected, compared to the single burst detected for FRB 180916.J0158+65 in this work.This means that the (flat-model) single-burst 0.5–10 keV energy limit for prompt emission fromFRB 180916.J0158+65, 1.6 × 1045 erg, is only ∼ 3× more constraining than the combined limit forFRB 121102, 4 × 1045 erg, which was derived under the assumption that an X-ray burst of similarfluence was emitted near the time of each radio burst.

The NH values assumed in the above calculations are the same as those taken for FRB 121102(Scholz et al. 2017), but may not be applicable for FRB 180916.J0158+65. From the DM budget pre-sented by Marcote et al. (2020) and the DM–NH relation from He et al. (2013), we can estimate whatthe NH towards FRB 180916.J0158+65 could be. The total DM measured for FRB 180916.J0158+65is 349 pc cm−3. Assuming the intergalactic medium (IGM) does not contribute significantly to NH,we subtract the IGM contribution to the DM, determined from the DM–z relation (Inoue 2004),34 pc cm−3. This leaves a Milky Way plus host DM of 291 pc cm−3, which from the DM–NH relationroughly corresponds to NH = 1022 cm−2, as used above. The high NH value, 1024 cm−2, was used inScholz et al. (2017) to simulate extreme X-ray absorption local to the source due to a high ratio ofatomic metals to free electrons, which could occur in a decades-old supernova remnant (Metzger et al.2017). However, Chawla et al. (2020) argues against such a young remnant for FRB 180916.J0158+65because of their recent detection of FRB 180916.J0158+65 at 300 MHz. This detection limits thesize, and thus age, of a remnant due to the requirement that the environment is optically thin tofree-free absorption at 300 MHz. As such, we consider this highly absorbed scenario unlikely forFRB 180916.J0158+65, though still consider it here for comparison to past limits on FRB 121102.

11

100 101 102 103

Photon Energy (keV)

1043

1044

1045

1046

1047

1048

1049

Bur

st e

nerg

y (e

rg)

Models (see Table 2)kT=10; NH=1022 cm 2

kT=10; NH=1024 cm 2=0.5; NH=1022 cm 2

=0.5; NH=1024 cm 2=2; NH=1022 cm 2

=2; NH=1024 cm 2

Models (see Table 2)kT=10; NH=1022 cm 2

kT=10; NH=1024 cm 2=0.5; NH=1022 cm 2

=0.5; NH=1024 cm 2=2; NH=1022 cm 2

=2; NH=1024 cm 2

Figure 2. Limits on the energy of X-ray and gamma-ray bursts at the time of radio bursts from

FRB 180916.J0158+65 (in black; this work) and FRB 121102 (in blue; from Scholz et al. 2017). The limits in

the 0.5–10 keV range are from Chandra, and in the 10–100 keV range are from Fermi/GBM. Dashed and solid lines

show the 5-σ upper limits as a function of X-ray photon energy, at the time of a single radio burst and stacking

those limits (see Section 4.1), respectively. The dot-dashed lines show different burst spectra that are photoelectrically

absorbed, assuming NH = 1022 cm−2, plotted at their 0.5–10 keV fluence limits that result from a stacked search of the

times of the radio bursts. The dotted lines show the same spectral models but with NH=1024 cm−2 to show the effects

of possible heavy absorption local to the source. Orange lines represent a blackbody model with kT = 10 keV, green

curves shows a cutoff power-law model with Γ = 0.5 and Ecut = 500 keV, and the grey curves show a soft power-law

with Γ = 2 in order to illustrate how different underlying spectra affect the interpretation of the X-ray observations.

12

Our burst limits can be compared to those placed for FRB 180916.J0158+65 using other telescopes.Tavani et al. (2020a) place a 3-σ persistent 0.3–10 keV X-ray flux of 5.5 × 10−14 erg s−1 cm−2 using10 ks of Swift/XRT observations during active periods of FRB 180916.J0158+65. For our corre-sponding limit we use a more stringent 5-σ confidence interval. Our 3-σ limit, however, would be4 × 10−15 erg s−1 cm−2, just over an order of magnitude deeper than the Swift/XRT limit. UsingINTEGRAL/IBIS, Panessa et al. (2020) place 3-σ upper limits on the 28–80 keV gamma-ray fluxof 3.4 × 10−8 erg s−1 cm−2 for 100-ms-long bursts at anytime during the INTEGRAL observations.This is very similar to the 10–100 keV Fermi/GBM limit placed here on gamma-ray emission at thetime of radio bursts (translated to a 3-σ limit on flux it is 3 × 10−8 erg s−1 cm−2).

4.2. Comparison to FRB Models

We can compare our X-ray and gamma-ray energy limits to the energy emitted by the 2004 giantflare of SGR 1806−20, the most energetic event detected from a Galactic magnetar. Though mostinteresting in the context of the magnetar model, this event is the most luminous transient event yetdetected in our Galaxy, so is therefore interesting in a model-agnostic context as well. The brightonset of the flare had a spectrum similar to that of our canonical giant flare model, an isotropicgamma-ray luminosity of ∼ 1047 erg s−1 (measured in the ∼ 20 keV–10 MeV band; Mazets et al.2005; Palmer et al. 2005), and a duration of ∼ 100 ms. This gives an emitted energy in a 10 keV–1 MeV band of ∼ 1046 erg. Our gamma-ray extrapolated isotropic energy limit for the giant-flare-likecutoff power-law model in Table 2 is still an order of magnitude higher than this energy emitted bySGR 1806−20. Further, Galactic magnetar activity includes much fainter events. The giant flaresfrom magnetars SGR 0526−66 and SGR 1900+14 had peak luminosities of 1044−45 erg s−1, over100× lower than the SGR 1806−20 giant flare. Short X-ray bursts from magnetars span far fainterluminosities (∼ 1036 − 1043 erg s−1; e.g., Gogus et al. 1999, 2000; Scholz & Kaspi 2011).

For the synchrotron blast wave model of FRBs, Metzger et al. (2019) and Margalit et al. (2019)predict an expected maximum fluence for a gamma-ray flare of ∼ 10−13 − 10−12 erg cm−2 forFRB 180916.J0158+65. This is far below the detection threshold of either our extrapolated X-ray lim-its (which would depend heavily on what the spectrum of the gamma-ray flare would be in the soft X-ray band) or our Fermi limits. The above shows that although the distance to FRB 180916.J0158+65is low for an FRB, it is still much too distant to probe the energies expected for magnetar-like activity.

The discovery of a 16.35-day periodicity in the radio burst activity of FRB 180916.J0158+65(CHIME/FRB Collaboration et al. 2020) has recently led to models in which the source — stillin many models a neutron star — is in an orbit or precessing. However, the current models do notclearly predict X-ray or gamma-ray emission that would be detectable using current instruments,given the distance to FRB 180916.J0158+65. For example, Mottez et al. (2020) describe a situationin which the relativistic wind of a pulsar or magnetar impinges on an orbiting planetary companion,creating an Alfven wing that if viewed downstream could be a source of FRBs. Given that thisscenario does not require powerful flares from the neutron star itself, observable X-ray emission atthe distance of FRB 180916.J0158+65 is not expected. Ioka & Zhang (2020) present a binary ‘comb’model in which FRBs are produced when the magnetosphere of a neutron star interacts with the windof a massive stellar companion, but make no specific predictions for the brightness of high-energyemission. Levin et al. (2020) note that a hyper-active magnetar that is driven by fast ambipolardiffusion in the core is expected to precess freely with a period of hours to weeks. This could explain

13

the periodicity of observed burst activity, but there is no reason to think that the magnetar flaresthemselves would be intrinsically brighter or dimmer compared to those we have considered above.

Persistent X-ray emission from FRB sources could arise from a pulsar wind nebula (if the FRBsource is a rotation or magnetically powered pulsar). We therefore compare our limit to the X-ray luminosity of the Crab Nebula, 1037 erg s−1. This is three orders-of-magnitude lower than ourpersistent X-ray luminosity limit of 2 × 1040 erg s−1. We can also compare our X-ray luminositylimit to the luminosities of the brightest X-ray sources. It is comparable to the luminosities oflow-luminosity active galactic nuclei (Terashima & Wilson 2003), bright high-mass X-ray binaries(Sazonov & Khabibullin 2017), and ultraluminous X-ray sources (Earnshaw et al. 2019). For all ofthese sources, their luminosity distributions extend well below our limit so we cannot rule out any suchassociation with the source of FRB 180916.J0158+65. However, it shows that future observationsof FRBs closer than FRB 180916.J0158+65 have the potential to make a detection if any of theseobjects are associated with the source.

Note that when translating our flux and fluence limits here to limits on luminosity or energy, weassume an isotropic energy release. If the high-energy emission from an FRB source is beamed, theenergy emitted would of course be lower as its emitted over a narrower solid angle.

For both prompt and persistent emission, we are only just beginning to probe the brightest ofpossible counterparts to repeating FRBs. Even for the closest sources, say at < 100 Mpc, ruling outhigh-energy activity from most models, such as that expected from a magnetar, is challenging. It is,however, important to place the most stringent possible limits for closer sources, in case there aremuch more energetic counterparts to repeating FRBs.

Late in the preparation of this work, we became aware of the works of Pilia et al. (2020) andTavani et al. (2020b) where limits were placed on the high-energy emission of FRB 180916.J0158+65during its active phases using XMM-Newton, Swift/XRT and AGILE. The deep XMM-Newton limitsplaced on the X-ray emission by Pilia et al. (2020) at the time of radio bursts using are similar toours placed here with Chandra. The AGILE limits probe a higher energy range than we consideredhere with Fermi/GBM. The persistent X-ray emission limits from Swift (Tavani et al. 2020b) andXMM-Newton (Pilia et al. 2020) are consistent with those we place here.

14

ACKNOWLEDGMENTS

We thank the Dominion Radio Astrophysical Observatory, operated by the National ResearchCouncil Canada, for gracious hospitality and useful expertise. The CHIME/FRB Project is fundedby a grant from the Canada Foundation for Innovation 2015 Innovation Fund (Project 33213), as wellas by the Provinces of British Columbia and Quebec, and by the Dunlap Institute for Astronomyand Astrophysics at the University of Toronto. Additional support was provided by the CanadianInstitute for Advanced Research (CIFAR), McGill University and the McGill Space Institute via theTrottier Family Foundation, and the University of British Columbia. The Dunlap Institute is fundedby an endowment established by the David Dunlap family and the University of Toronto. Research atPerimeter Institute is supported by the Government of Canada through Industry Canada and by theProvince of Ontario through the Ministry of Research & Innovation. The National Radio AstronomyObservatory is a facility of the National Science Foundation operated under cooperative agreementby Associated Universities, Inc. This work is based on observations with the 100-m telescope ofthe MPIfR (Max-Planck-Institut fur Radioastronomie) at Effelsberg. We thank the DSN schedulingteam and the Goldstone Deep Space Communication Complex (GDSCC) and the Madrid Deep SpaceCommunication Complex (MDSCC) staff for scheduling and carrying out the DSN observations. Aportion of this research was performed at the Jet Propulsion Laboratory, California Institute ofTechnology and the Caltech campus, under a Research and Technology Development Grant througha contract with the National Aeronautics and Space Administration. U.S. government sponsorshipis acknowledged.

A.B.P. acknowledges support by the Department of Defense (DoD) through the National DefenseScience and Engineering Graduate (NDSEG) Fellowship Program and by the National Science Foun-dation (NSF) Graduate Research Fellowship under Grant No. DGE-1144469. B.M. acknowledges sup-port from the Spanish Ministerio de Economıa y Competitividad (MINECO) under grant AYA2016-76012-C3-1-P. D.M. is a Banting Fellow F.K. is supported by the Swedish Research Council. FRBresearch at UBC is supported by an NSERC Discovery Grant and by the Canadian Institute forAdvanced Research. J.W.T.H. acknowledges funding from an NWO Vici fellowship L.G.S. is a Lise-Meitner independent research group leader and acknowledges support from the Max Planck Society.M.B. is supported by an FRQNT Doctoral Research Award. P.C. is supported by an FRQNTDoctoral Research Award. P.S. is a Dunlap Fellow and an NSERC Postdoctoral Fellow. B.M.G.acknowledges the support of the Natural Sciences and Engineering Research Council of Canada(NSERC) through grant RGPIN-2015-05948, and of the Canada Research Chairs program. V.M.K.holds the Lorne Trottier Chair in Astrophysics & Cosmology and a Canada Research Chair andreceives support from an NSERC Discovery Grant and Herzberg Award, from an R. Howard WebsterFoundation Fellowship from the Canadian Institute for Advanced Research (CIFAR), and from theFRQNT Centre de Recherche en Astrophysique du Quebec. W.A.M, T.A.P, and C.J.N acknowl-edge support by the Jet Propulsion Laboratory’s Spontaneous Concept Research and TechnologyDevelopment program. Z.P. is supported by a Schulich Graduate Fellowship.

REFERENCES

Agarwal, D., Aggarwal, K., Burke-Spolaor, S.,Lorimer, D. R., & Garver-Daniels, N. 2019,arXiv e-prints, arXiv:1902.06343.https://arxiv.org/abs/1902.06343

An, H., Kaspi, V. M., Beloborodov, A. M., et al.

2014, ApJ, 790, 60

15

Atwood, W. B., Abdo, A. A., Ackermann, M.,et al. 2009, ApJ, 697, 1071,doi: 10.1088/0004-637X/697/2/1071

Bannister, K. W., Deller, A. T., Phillips, C., et al.2019, Science, 365, 565,doi: 10.1126/science.aaw5903

Beloborodov, A. M. 2017, ApJL, 843, L26,doi: 10.3847/2041-8213/aa78f3

Chatterjee, S., Law, C. J., Wharton, R. S., et al.2017, Nature, 541, 58.https://arxiv.org/abs/1701.01098

Chawla, P., Andersen, B. C., Bhardwaj, M., et al.2020, arXiv e-prints, arXiv:2004.02862.https://arxiv.org/abs/2004.02862

CHIME/FRB Collaboration, Amiri, M., Bandura,K., et al. 2018, ApJ, 863, 48,doi: 10.3847/1538-4357/aad188

CHIME/FRB Collaboration, :, Andersen, B. C.,et al. 2019, ApJL, 885, L24.https://arxiv.org/abs/1908.03507

CHIME/FRB Collaboration, Amiri, M.,Andersen, B. C., et al. 2020, Nature.https://arxiv.org/abs/2001.10275

Cordes, J. M., & Chatterjee, S. 2019, ARA&A, 57,417, doi: 10.1146/annurev-astro-091918-104501

Cordes, J. M., & Lazio, T. J. W. 2002, arXivpreprint astro-ph/0207156

Cunningham, V., Cenko, S. B., Burns, E., et al.2019, ApJ, 879, 40,doi: 10.3847/1538-4357/ab2235

de Jager, O. C., Raubenheimer, B. C., &Swanepoel, J. W. H. 1989, A&A, 221, 180

Earnshaw, H. P., Roberts, T. P., Middleton, M. J.,Walton, D. J., & Mateos, S. 2019, MNRAS,483, 5554, doi: 10.1093/mnras/sty3403

Evans, W. D., Klebesadel, R. W., Laros, J. G.,et al. 1980, ApJL, 237, L7

Fruscione, A., McDowell, J. C., Allen, G. E., et al.2006, in Society of Photo-OpticalInstrumentation Engineers (SPIE) ConferenceSeries, Vol. 6270, Society of Photo-OpticalInstrumentation Engineers (SPIE) ConferenceSeries, doi: 10.1117/12.671760

Gajjar, V., Siemion, A. P. V., Price, D. C., et al.2018, ApJ, 863, 2,doi: 10.3847/1538-4357/aad005

Gogus, E., Woods, P. M., Kouveliotou, C., et al.1999, ApJL, 526, L93

Gogus, E., Woods, P. M., Kouveliotou, C., et al.2000, ApJL, 532, L121

He, C., Ng, C.-Y., & Kaspi, V. 2013, TheAstrophysical Journal, 768, 64

Hotan, A. W., van Straten, W., & Manchester,R. N. 2004, PASA, 21, 302,doi: 10.1071/AS04022

Hurley, K., Cline, T., Mazets, E., et al. 1999,Nature, 397, 41, doi: 10.1038/16199

Hurley, K., Boggs, S. E., Smith, D. M., et al. 2005,Nature, 434, 1098, doi: 10.1038/nature03519

Inoue, S. 2004, MNRAS, 348, 999Ioka, K., & Zhang, B. 2020, arXiv e-prints,

arXiv:2002.08297.https://arxiv.org/abs/2002.08297

Kaspi, V. M., & Beloborodov, A. M. 2017, AnnualReview of Astronomy and Astrophysics, 55, 261

Kong, A. K. H., Li, K. L., Hui, C. Y., & Yu, H. F.2020, The Astronomer’s Telegram, 13589, 1

Kraft, R. P., Burrows, D. N., & Nousek, J. A.1991, ApJ, 374, 344, doi: 10.1086/170124

Levin, Y., Beloborodov, A. M., & Bransgrove, A.2020, arXiv e-prints, arXiv:2002.04595.https://arxiv.org/abs/2002.04595

Lin, L., Gogus, E., Baring, M. G., et al. 2012,ApJ, 756, 54

Lyubarsky, Y. 2014, MNRAS, 442, L9,doi: 10.1093/mnrasl/slu046

Majid, W. A., Pearlman, A. B., Dobreva, T.,et al. 2017, ApJL, 834, L2,doi: 10.3847/2041-8213/834/1/L2

Marcote, B., Nimmo, K., Hessels, J. W. T., et al.2020, Nature, 577, 190,doi: 10.1038/s41586-019-1866-z

Margalit, B., Metzger, B. D., & Sironi, L. 2019,arXiv e-prints, arXiv:1911.05765.https://arxiv.org/abs/1911.05765

Mazets, E. P., Cline, T. L., Aptekar, R. L., et al.2005, ArXiv Astrophysics e-prints

Meegan, C., Lichti, G., Bhat, P. N., et al. 2009,ApJ, 702, 791

Metzger, B. D., Berger, E., & Margalit, B. 2017,ApJ, 841, 14, doi: 10.3847/1538-4357/aa633d

Metzger, B. D., Margalit, B., & Sironi, L. 2019,arXiv e-prints.https://arxiv.org/abs/1902.01866

Mottez, F., Voisin, G., & Zarka, P. 2020, arXive-prints, arXiv:2002.12834.https://arxiv.org/abs/2002.12834

Palmer, D. M., Barthelmy, S., Gehrels, N., et al.2005, Nature, 434, 1107

16

Panessa, F., Savchenko, V., Ferrigno, C., Bazzano,A., & Ubertini, P. 2020, The Astronomer’sTelegram, 13466, 1

Pearlman, A. B., Majid, W. A., & Prince, T. A.2019, Advances in Astronomy, 2019, 6325183,doi: 10.1155/2019/6325183

Pearlman, A. B., Majid, W. A., Prince, T. A.,Kocz, J., & Horiuchi, S. 2018, ApJ, 866, 160,doi: 10.3847/1538-4357/aade4d

Petroff, E., Hessels, J. W. T., & Lorimer, D. R.2019, A&A Rv, 27, 4,doi: 10.1007/s00159-019-0116-6

Pilia, M., Burgay, M., Possenti, A., et al. 2020,arXiv e-prints, arXiv:2003.12748.https://arxiv.org/abs/2003.12748

Platts, E., Weltman, A., Walters, A., et al. 2018,arXiv e-prints.https://arxiv.org/abs/1810.05836

Prochaska, J. X., Macquart, J.-P., McQuinn, M.,et al. 2019, Science, 366, 231,doi: 10.1126/science.aay0073

Ransom, S. 2011, PRESTO: PulsaR Explorationand Search TOolkit, Astrophysics Source CodeLibrary. http://ascl.net/1107.017

Ravi, V., Catha, M., D’Addario, L., et al. 2019,Nature, 572, 352,doi: 10.1038/s41586-019-1389-7

Sazonov, S., & Khabibullin, I. 2017, MNRAS, 466,1019, doi: 10.1093/mnras/stw3113

Scholz, P., & Kaspi, V. M. 2011, ApJ, 739, 94

Scholz, P., Bogdanov, S., Hessels, J. W. T., et al.

2017, ApJ, 846, 80,

doi: 10.3847/1538-4357/aa8456

Spitler, L. G., Scholz, P., Hessels, J. W. T., et al.

2016, Nature, 531, 202.

https://arxiv.org/abs/1603.00581

Tavani, M., Verrecchia, F., Casentini, C., et al.

2020a, The Astronomer’s Telegram, 13462, 1

—. 2020b, arXiv e-prints, arXiv:2004.03676.

https://arxiv.org/abs/2004.03676

Tendulkar, S. P., Kaspi, V. M., & Patel, C. 2016,

ApJ, 827, 59, doi: 10.3847/0004-637X/827/1/59

Terashima, Y., & Wilson, A. S. 2003, ApJ, 583,

145, doi: 10.1086/345339

Thornton, D., Stappers, B., Bailes, M., et al.

2013, Science, 341, 53.

https://arxiv.org/abs/1307.1628

Verner, D. A., Ferland, G. J., Korista, K. T., &

Yakovlev, D. G. 1996, ApJ, 465, 487

Wilms, J., Allen, A., & McCray, R. 2000, ApJ,

542, 914

Yao, J. M., Manchester, R. N., & Wang, N. 2017,

ApJ, 835, 29. https://arxiv.org/abs/1610.09448